We have so far established that an increase in deposit insurance coverage does not cause an increase in risk taking, disproving one of the more commonly cited potential pitfalls in providing
9In this specification, an interaction termD
af terDcrisisshould be added. However, the treatment period and the
crisis period overlap (i.e. the crisis period is fully contained in the after treatment period), hence this interaction is collinear and therefore redundant.
Table 2.4: Results for crisis period specification (1) (2) VARIABLES γ1 t,i γt,i1 DT 0.0104*** 0.0104*** (0.00304) (0.00304) Daf ter -0.0223*** -0.0256*** (0.00334) (0.00350) DTDaf ter -0.00172 -0.00309 (0.00441) (0.00467) Dcrisis 0.00813 (0.00574) DTDcrisis 0.00343 (0.00687) Constant 0.0342*** 0.0342*** (0.00205) (0.00205) Observations 1,451 1,451 R-squared 0.073 0.082
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
deposit insurance or increasing its coverage. As theory predicts, the moral hazard effect of an increase in the coverage of deposit insurance is that deposit supply increases and depositors reduce their monitoring of bank behaviour. Increased deposit supply and laxer market discipline should, given a level of demand for deposit funding, decrease the interest rate on deposits.
To test whether an increase in deposit insurance coverage affects the deposit supply and market discipline we use the difference-in-differences estimation. Theoretical prediction states that an in- crease in deposit insurance coverage decreases deposit rates as a result of the deposit supply rise. We therefore run the following regression:
it,i=β0+β1DT +β2Daf ter+β3DTDaf ter+it (2.4)
whereit,irefers to the deposit rate of bankiin periodt. As before, the coefficient of interest isβ3,
which provides the estimate of the treatment effect of an increase in deposit insurance coverage on the deposit interest rates. Since we do not have data on deposit rates for new deposit transactions, following Schmukler (2001), we compute the implicit interest rate, i.e. the ratio of total interest expense on domestic deposits over total domestic deposits for each bank each year. In using implicit deposit rates, we may be overlooking deposit composition effects. It cannot be excluded that deposits with different maturities are subject to different rates. Averaging interest expenses across deposit classes may therefore be an issue.
Figure 2.3 plots the means of the implicit deposit interest rates over the horizon for the control group and the treatment group as defined in the previous section. The figure provides several insights. Firstly, the two series clearly exhibit a common trend in the pre-treatment periods, suggesting that the series are suitable to perform the estimation on. Secondly, the lack of divergence in the post- treatment periods already hints at the conclusions which are confirmed by our estimation results in Table 2.5.
Figure 2.3: Means of the implicit deposit rates for all the FDIC insured banks and the Massachusetts state chartered banks.
Table 2.5: Effect of deposit insurance on deposit rates
(1)
VARIABLES deposit rate
DT 0.178*** (0.0488) Daf ter -1.131*** (0.0521) DTDaf ter -0.104 (0.0670) Constant 2.167*** (0.0375) Observations 2,028 R-squared 0.393
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The coefficient of interest,DTDaf ter, corresponding toβ3 in equation 2.4, is statistically insignif-
icant, indicating that the increase in deposit insurance coverage does not increase deposit supply which would result in lower interest rate.
This result provides an explanation for the result of no response of bank risk taking following an increase in deposit insurance coverage, presented in the previous section. After in increase in deposit
Figure 2.4: Deposits of the Massachusetts state chartered banks.
insurance coverage, the depositors do not increase their supply of funding to the bank which does not result in laxer market discipline. Our results are in line with those of Schmukler (2001).
2.6
Robustness
2.6.1
Indirect treatment effect on the non-treated through the deposit
supply
Our results rely on the assumptions of matching and the difference-in-differences approaches. We have established a treatment and a control group of banks that are similar in observables. However, to assure their validity, we need to make sure that the control group is unaffected by the treatment. The control group in our experiment is comprised of banks whose deposit insurance is unlimited, which implies that they are unaffected by the country-wide increase in the national coverage limit from $100,000 to $250,000. Although we see no way for the control group to be affected directly by the treatment, there remains a plausible indirect effect through deposit supply. In this case depositors in Massachusetts state chartered banks, whose choice of bank was based on the fact that their funds are insured, could withdraw their funds in excess of $100,000 and opt for a bank which now also offers insured deposits but is more convenient along some other dimension.
Figure 2.4 shows the average total balance of deposits for each bank, and how they are split between those which lie above the insurance limit and those below. The two kinks in the amount of deposits above and below the limit correspond to the change in policy. Note that with the increase in coverage, deposits between $100,000 and $250,000 were above the limit before the change, and are below the limit after, so the amount of deposits below the limit increases merely by accounting. The same logic can be applied to deposits above the limit. However, the total amount of deposits
does not deviate from its trend. It can thus be concluded that the banks in the control group did not experience any deposit flight due to the policy change.
2.6.2
Results for different treatment dates
As stated before, the increase in deposit insurance that happened in 2008 was intended as a temporary measure. However, it was made permanent in July 2010. In order to make sure that the absence of a treatment effect is not due to the temporary nature of this measure, we estimate the same model with 2010 as a treatment date.
Furthermore, in order to ensure that the lack of effect is not due to anticipation effects, we use 2007 as a treatment year, to account for the case that some information regarding an increase in deposit insurance coverage circulated prior to the implementation.
Table 2.6: Different treatment dates
(1) (2)
VARIABLES Treatment in 2007 Treatment in 2010
DT 0.00936*** 0.00989*** (0.00334) (0.00270) Daf ter -0.0175*** -0.0254*** (0.00332) (0.00375) DTDaf ter -4.93e-05 -0.00160 (0.00448) (0.00466) Constant 0.0325*** 0.0318*** (0.00227) (0.00187) Observations 1,451 1,451 R-squared 0.040 0.099
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 2.6 shows that the treatment effect is still insignificant for these treatment dates. Again, we find no evidence of a moral hazard problem that would result in higher risk taking by banks after an increase in deposit insurance coverage.